We investigated the dose-response characteristics of glucose-induced insulin release and the influence of hyperglycaemia on arginine-induced insulin secretion in eight non-obese subjects with NIDDM and in eight non-diabetic volunteers. Plasma C-peptide levels, achieved during 60 min hyperglycaemic clamps with and without the infusion of a primed continuous infusion of arginine (infusion rate 15 mg kg-1 min-1) during the last 30 min, were analysed with a modified Michaelis-Menten equation. The insulin secretory capacity (Vmax) for glucose-stimulated insulin release showed a trend towards a negative correlation with the fasting blood glucose in the NIDDM subjects (r = 0.68, P = 0.6); it was lower than the Vmax of non-diabetic controls (2.2 +/- 0.2 vs 4.2 +/- 0.4 nmol l-1 respectively; P less than 0.001). The ED50 (half maximal stimulating blood glucose concentration) of the second-phase glucose-stimulated insulin release (determined from the plasma C-peptide levels at 60 min) was not significantly different from the ED50 of the controls (11.9 +/- 0.8 vs 13.3 +/- 1.9 mmol l-1 respectively; P greater than 0.2). Combined glucose-arginine stimulation significantly increased insulin release. The Vmax for both phases were significantly lower in NIDDM patients than in controls (2.3 +/- 0.2 vs 5.0 +/- 0.9 and 3.8 +/- 0.5 vs 8.5 +/- 0.9 nmol l-1 respectively; P less than 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)